Paleomagnetism is our best tool for mapping the Earth in the deep past. One difficulty with it, though, is that it gives ambiguous results. Researchers can only tell that a rock's magnetic field is pointing to one of the magnetic poles, which may be north or south. If a continent has been somewhere where the choice of poles is not clear, it has two possible pasts—one in the northern hemisphere and one in the southern hemisphere—and no obvious way to tell which one is correct. We have figured out clever ways to get around that.
Consider a drifting continent whose rocks have a good record of paleopoles and paleolatitudes. The record, going backward in time, shows that the continent moved to the equator, then away from the equator. Did it cross the equator into the other hemisphere, or did it back off instead? Magnetic data cannot tell us the answer, even when the record is perfect. The problem is worse when the record is spotty, with gaps of tens of millions of years or longer.
There are other types of evidence we can combine with paleomagnetic data. For some continents, we can figure out where they once fit against other continents. That can be done with fossils—if identical or closely related fossils occur on two separate continents, like Europe and North America, we can infer that they were once joined, perhaps in a supercontinent. That kind of evidence has a long pedigree, although useful fossils are limited to Cambrian time and later.
We can do the same with geologic features, like large plutons or major igneous dikes that can be mapped across the gaps between continents. We can reason that the rock must have taken its position when the continents were joined. If one of those continents was certifiably in the northern or the southern hemisphere, then the other one must have been too.
With rocks from the 4-billion-year span of Precambrian time, matching dikes and plutons is the best method we have, but it's not always possible. Even when it's possible, the evidence that certifies at least one continent to be in one particular hemisphere may be missing.
The Coriolis Trick
There's one good way past the problem, although it takes a lot of careful work and luck (which is true of all geology, come to think of it). It's based on the one sure difference between the northern and southern hemispheres: the Coriolis force. That's the effect the Earth's rotation has on currents, whether they're ocean currents or atmospheric winds. The Coriolis force turns any motion on a sphere's surface into a clockwise loop, like the Gulf Stream current in the Atlantic Ocean or the Kuroshio in the Pacific—but only in the northern hemisphere. In the southern hemisphere it works counterclockwise. It also means that in the tropics near the equator, the prevailing winds (the trade winds) always blow from the east.
Now picture this: If you're in the northern hemisphere facing the north magnetic pole, the trade winds come from your right. If you're in the southern hemisphere facing the south magnetic pole, they blow from your left. By the same token, the ocean currents will be flowing in opposite directions. Find a way to test this in the geologic record, and you can put yourself in the correct hemisphere.
John Grotzinger is famous today for being chief scientist of the Mars Curiosity rover mission. But in the mid-1980s, Grotzinger was a doctoral student in Canada studying the Slave craton, a chunk of the Canadian Shield some 2 billion years old that had been somewhere in the tropics around 1900 million years ago, oriented upside-down with respect to today. In the ancient limestones along its west side, stromatolites were strongly carved by the trade winds, indicating their direction. The presence of a wide limestone shelf showed that it had been on the upwind side of the craton next to a warm ocean, like the Philippines or the Caribbean islands of today. That could only have happened in the northern hemisphere. Had the craton been turned around and south of the equator instead, those same winds would have been dry offshore trade winds and the ocean cold, like the South American west coast of today (which is decidedly not limestone country).
The Tectonic Style Model
In 1993 Grotzinger and Paul Hoffman published a more speculative paper in Geology with the cryptic title "Orographic precipitation, erosional unloading, and tectonic style." The Slave craton again was their test case, and this time they looked at how geography would have influenced geology. Two ancient mountain belts were pushed up against the Slave craton from opposite sides about 90 million years apart, first the Thelon orogen and then the Wopmay orogen. These formed, as all orogens do, from collisions with other plates.
The side of the mountains facing the trade winds would have captured most of the rain—that's the "orographic precipitation" in the title. That rainfall (and snowfall in the high country) would have vigorously eroded the mountains, allowing the lithosphere underneath to rise as the eroded weight was removed—that's the "erosional unloading" part. As this continued and the plate continued to warp with the changes, the orogen and the rocks that formed around it would take on a lopsided appearance (today the Tibetan Plateau is undergoing similar changes). Their model made sense of the different histories, structures and appearances—the "tectonic styles"—of the Thelon and Wopmay orogens and the evidence they left on the Slave craton. Knowing north from south was essential to Grotzinger and Hoffman's work.
The paper was influential and has been cited many times since 1993. Conceivably, other well-preserved Precambrian cratonic rock records could fit this model and help tell us ancient north from ancient south. But so far, the model doesn't seem to have worked anywhere else, because it requires a lot of evidence to work well and, for the most part, Precambrian cratons have lost that kind of evidence. But we're still mapping and still learning, and all sorts of half-forgotten papers find themselves unexpectedly relevant again.